Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
1
INSTITUTO NACIONAL DE PESQUISAS DA AMAZÔNIA – INPA
PROGRAMA DE PÓS-GRADUAÇÃO EM CIÊNCIAS DE FLORESTAS
TROPICAIS –PPG/CFT
Herbivory and litter production in response to nutrient addition in
“Terra Firme” forest in Central Amazon
Anna Carolina Martins Moraes
Manaus, Amazonas
Junho, 2018
2
Anna Carolina Martins Moraes
Herbivory and litter production in response to nutrient addition in
“Terra Firme” forest in Central Amazon
Orientador: Dr. Carlos Alberto Quesada
Coorientadora: Dra. Kelly Marie Andersen
Dissertação apresentada ao Instituto
Nacional de Pesquisas da Amazônia como parte dos
requisitos para a obtenção do título de Mestre em
Ciências de Florestas Tropicais.
Manaus, Amazonas
Junho, 2018
3
M827h Moraes, Anna Carolina Martins
Herbivory and litter production in response to nutrient addition in “Terra
Firme” forest in Central Amazon – Herivoria e produção de serapilheira em
resposta à adição de nutrientes em floresta de terra firma na Amazônia Central/
Anna Carolina Martins Moraes. --- Manaus: [s.l.], 2018. 63 f.
Dissertação (Mestrado- Programa de Pós Graduação em Ciências
Florestas Tropicais) – Coordenação do Programa de Pós-Graduação,
INPA, 2018.
1.Amazônia. 2. Serapilheira. 3. Ciclagem de nutrientes. 4. Herbivoria.
5. Fertilização. I. Quesada, Carlos Alberto Nobre, oriente. II.
Andersern, Kelly, Marie, coorient. III. Título.
CD 333.75
Sinopse:
O foco do trabalho é entender o papel das limitações nutricionais na
produtividade do dossel, perdas herbívoras e investimentos em compostos secundários
em escala ecossistêmica na Amazônia Central. Nossos resultados
sugerem que a produção de serapilheira, folhas e herbivoria apresentaram influência com rápida resposta inicial à adição de nutrientes.
Palavras-chave: Amazônia, serapilheira, cicla gem de nutrientes, herbivoria,
fertilização.
4
Sumário
Resumo ...................................................................................................................................................... 11
Abstract ..................................................................................................................................................... 12
1. Introdução ......................................................................................................................................... 13
1.1. Introduction ................................................................................................................................ 18
2. Objective ........................................................................................................................................... 22
2.1. General Objective ....................................................................................................................... 22
2.2. Specific Objectives ..................................................................................................................... 22
3. Hypotheses ........................................................................................................................................ 22
4. Methods ............................................................................................................................................. 23
4.1. Location and characterization of the study area ......................................................................... 23
4.3. Calculations ................................................................................................................................ 26
4.4. Chemical analysis ....................................................................................................................... 27
4.5. Data analysis .............................................................................................................................. 29
5. Results ............................................................................................................................................... 30
5.1. Patterns of litterfall fractions with treatment .............................................................................. 30
5.2. Leaf Mass Area .......................................................................................................................... 32
5.3. Nutrient content in leaf litter ...................................................................................................... 32
5.4. Leaf litter nutrient inputs ............................................................................................................ 33
5.5. Defence Compounds .................................................................................................................. 35
5.6. Leaf Area and Biomass consumed by Herbivory ....................................................................... 35
5.7. Insect Frass Biomass .................................................................................................................. 36
5.8. Insect Frass Nutrient content ...................................................................................................... 37
5.9. Frass Biomass Nutrient dynamics .............................................................................................. 37
5.10. Interactions between leaf traits variables and herbivory frass .................................................. 38
6. Discussion ......................................................................................................................................... 43
6.1. Short-term responses to fertilization ..................................................................................... 43
6.2. Effect of nutrient addition in litterfall production ................................................................. 43
6.3. Effect of nutrient addition in foliar traits: LMA .................................................................... 45
6.4. Effect of nutrient addition in foliar traits: Litterfall nutrient concentrations ......................... 45
6.5. Effect of nutrient addition in foliar traits: Defense compounds ............................................ 47
6.6. Effect of Nutrient Addition in Leaf Herbivory...................................................................... 47
6.7. Effect of nutrient addition in herbivory: Insect frass biomass and nutrient content. .................. 48
6.8. Effect of Nutrient Addition in the interactions between foliar traits and frass herbivory .......... 49
6.9. Limitations of the study ......................................................................................................... 50
7. Conclusion ......................................................................................................................................... 51
5
7.1. Conclusão ................................................................................................................................... 51
8. Bibliographic references ................................................................................................................... 52
Summary of Figures
FIGURE 1. Map of BDFFP (Biological Dynamics of Forest Fragments Project) research area.
Yellor area indicates the location of the AFEX research site, named forest reserve “Km 41”. Source: E.M.
e Fernandes T.L.N. ......................................................................................................................................23
FIGURE 2.Example of the position diagram of plots inside one block with trails between the eight
plots in blue dotted line, and the treatments Control, Nitrogen (N), Phosphorus (P), Cations (CATIONS),
N+P, N+CATIONS, P+CATIONS, N+P+CATIONS. ................................................................................24
FIGURE 3.Picture of the plot design of 50 m² in grey, with the buffer zone of 10 m in red dotted
lines,the central area inside the plot with 30m² in green, and distribution of the five litter traps in yellow.
.....................................................................................................................................................................25
FIGURE 4.a) Littertrap installed in the plot; b) litterfall collection in one of the five littertraps in
the center plot area. Source: Andersen, K and Moraes, A.C.M. ..................................................................25
FIGURE 5. Example of leaves with and without herbivory used to calculate leaf area loss. a) is the
scanned leaf original area measured (with herbivory); b) is the calculation of the area missing projected in
the program Image J; c) is the leaf designed without herbivory. Source: K. Andersen e Moraes, A.C.M.
.....................................................................................................................................................................27
FIGURE 7. Mean fractions of fine litterfall Mg ha-1 year-1 (leaves in green color, fine wood in
orange, reproductive material in pink, other residues in blue, and frass in brown), by nutrient addition
treatments of Control (CTL), Nitrogen (N), Phosphorus (P), Cations (Cation), N+P (NP), N+Cations
(NCation), P+Cations (PCation), and N+P+Cations (NPCation). .............................................................. 33
FIGURE 8. Fine litterfall (Mg ha-1 year-1) with nutrient addition treatments of Control (CTL) in
red, Nitrogen (N) in yellow, Phosphorus (P) in light green, Cations (Cation) in dark green, N+P (NP) in
light blue, N+Cations (NCation) in turquoise blue, P+Cations (PCation) in lilac, and N+P+Cations
(NPCation) in pink. The x axis is counting the collection date over time (july/2017 – february/2018). .... 34
FIGURE 9. Cumulative leaf litterfall(g.m²)over time by treatments Control (CTL) in red, Nitrogen
(N) in yellow, Phosphorus (P) in light green, Cations (Cation) in dark green, N+P (NP) in light blue,
N+Cations (NCation) in turquoise blue, P+Cations (PCation) in lilac, and N+P+Cations (NPCation) in pink.
.................................................................................................................................................................... 34
FIGURE 10. Figure a show Leaf Mass Area (g.m²) with the factorial interaction of Nitrogen and
Cations. Showing that the LMA had a significant response with the presence of both Nitrogen and Cations.
6
Figure b show all treatments LMA pattern. ............................................................................................... 35
FIGURE 11. Quantity of K, Mg, Ca, Zn, Fe, Mn (g.kg-1) in all 8 treatments……………………35
FIGURE 12. Quantity of leaf nutrients (g.kg-1) of potassium (K) (figure a),Calcium (Ca) (figure
b), Magnesium (Mg) (figure c), Manganese (Mn) (figure d). The mean of control value in marked with
black dotted line.. ......................................................................................................... .............................. 36
FIGURE 13. Quantity of phosphorus (g.ha-1) (figure a and b), phosphorus response in the presence of
treatments P, and without the presence of Cations. Manganese (g.ha-1) (figure c) response with factorial
P treatment. ................................................................................................................................................. 37
FIGURE 14. Quantity of litter nutrient inputs (g.ha-1) of Calcium (Ca) (figure a), Magnesium (Mg)
(figure b), Potassium (K) (figure c). The x axis show the treatments Cations in red, Control in yellow (the
mean value in marked with black dotted line), Nitrogen (N) in light green, N+Cations in dark green, N+P
in light blue, N+P+Cations in turquoise blue, Phosphorus (P) in lilac , P+Cations in pink. ...................... 38
FIGURE 15.Percentage (%) of lignin (figure a), cellulose (figure b) and phenols (figure c)
remaning in the leaf litterfall in all eight treatments. The x axis show the treatments Cations in red, Control
in yellow (the mean value in marked with black dotted line), Nitrogen (N) in light green, N+Cations in dark
green, N+P in light blue, N+P+Cations in turquoise blue, Phosphorus (P) in lilac , P+Cations in pink. ... 38
FIGURE 16. Leaf are missing (%) (figure a), Leaf Biomass Herbivory (g.ha-1) (figure b) in all 8
treatments in the x axis. Showing the control mean value in marked with black dotted line. .................... 39
FIGURE 17. The Insect Frass Biomass (kg.ha-1), showing in all treatments, and the mean control
value in marked with black line.. ........................................................................................... ..................... 40
FIGURE 18. The Insect Frass Nutrient (g.kg-1), showing in figure “a” the nutrient calcium, figure
“b” the nutrient magnesium, figure “c” the nutrient potassium. The x axis being all treatments, the mean
control value in marked with black line. The figure “e” the factorial interaction of Phosphorus and Cations
significant response with the presence of both Phosphorus and Cations. ................................................... 40
FIGURE 19. Quantity of P, Mg, Ca and Mn in insect biomass frass (g.ha-1) with and without P,and
in all 8 treatments. , K . The x axis being all treatments and the mean control value in marked with black
line. ....................................................................................................................... ...................................... 41
FIGURE 20. PCA on leaf trait variables (nutrients phosphorus, mangnesium, calcium, potassium,
manganese (g.kg-1), LMA (g.m²), Cellulose, lignin and phenosl (%), based on means per plot. The two
principal components accounted for 51.2% of the variation data. The PCA included all eight nutrient
addition treatments of Control (red), Nitrogen (N) (yellow), Phosphorus (P) (light green), Cation (dark
green), N+P (NP) (light blue), N+Cation (NCation) (turquoise blue), P+Cation (PCation) (lilac),
N+P+Cation (NPCation) (pink). ............................................................................................... .................. 42
FIGURE 21. Simple correlation between concentration of the nutrients in insect frass biomass
input (g.ha-1) and leaf litter biomass input (g.ha-1) of potassium (figure a), magnesium (figure b), calcium
(figure c) and phosphorus (figure d) and figure e is the correlation between calcium (g.kg-1) and phenols.
.................................................................................................................................................................... 43
7
Summary of Tables
TABLE 1. List of treatments names in the factorial design used in AFEX fertilization. .............24
TABLE 2. Concentrations of the nutrients Nitrogen, Phosphorus and Cations in the treatments.
.................................................................................................................................................................... 27
TABLE 3. Table of variavles in treatments control. It contains the type of material, unit, number
of repetitions (N), mean value, standard deviation (SD), standard error (SE), and confidence Interval (CI).
.................................................................................................................................................................... 44
8
Para minha momis, papis, queridos irmãos Harley’s
E minha anja Eliana,
sei que a senhora está me vendo e me apoiando.
9
Agradecimentos
Primeiramente, agradeço à Deus por me dar as forças necessárias mesmo quando eu
achava que não havia nenhuma. Aos meus pais e irmãos, por me darem todo amor, coragem, apoio
incondicional e estrutura em todos os sentidos para conseguir realizar essa aventura de mestrado.
Aos meus incríveis orientadores Beto Quesada e Kelly Andersen, por toda a força e
ensinamentos no mundo da pesquisa, como reciprocidade de entender as nossas angústias,
transparência e esclarecimentos no mundo da ciência. Obrigada Beto, que só de aparecer no
corredor já nos acalmava com o sorriso meigo, abraço forte e segurança de que tudo daria certo.
Obrigada Kelly, que mesmo muitas vezes do outro lado do oceano, foi muito presente em todos
os momentos, sempre disposta a responder todas as nossas dúvidas, e muito motivadora sempre,
uma mulher realmente incrível e forte. Vocês me ensinaram que ciência não é sobre ser um gênio
e sim sobre ter esforço de querer aprender, e reconhecer que precisamos buscar conhecimento,
perseverança e amor no trabalho. Obrigada pela confiança na realização dessa pesquisa, por não
silenciar minhas ideias e também empoderar mulheres do nosso laboratório para a futuro de
pesquisas que nos espera a frente.
As minhas amigas lindas, sábias e sensatas, principalmente nos meus momentos de
dúvidas e questionamentos sobre a vida, Ju Lopes e Bruna Oliveira, obrigada por sempre estarem
comigo, principalmente na hora do almoço, desde a faculdade.
Todo esse infinito apoio, me deu motivação para passar 2/3 do meu dia no INPA sem nem
perceber o tempo passar, junto com a maioria das minhas companheiras do laboratório que
também faziam isso tranquilamente, Jé Rosa, Mandinha Longhi, e Nathylinda Martins
companheiras de sofrência, que não me deixaram desistir, me apoiando sempre, com muitos
abraços, ombro amigo e conversas infinitas, me pegando literalmente do chão e enxugando minhas
lágrimas. A Lalay Lugli, que me ajudou com muita sabedoria e sagacidade, mesmo também no
outro lado do mundo, não medindo esforços para ajudar no que quer que fosse. Ver vocês em ação
foi uma das minhas maiores motivações, ícones da ciência!
As minhas companheiras de triagem de serapilheira maravilhosas, ágeis, organizadas, e
engraçadíssimas, Adriana e Maria, sem o comprometimento de vocês eu não conseguiria fazer
todo esse experimento, obrigada por tudo.
Ao laboratório LTSP, por me ajudar grandemente nas análises químicas super difíceis e
cansativas, Gabi, Jonas, Roberta, Laura, Ed, e Raimundo. Seus conhecimentos e experiências são
valiosos para que façamos análises com o mínimo de erro possível, mesmo eu não colaborando
tanto com minha falta de experiência. Obrigada pela paciência e por nos ajudar sempre.
10
Aos amigos do laboratório beergeoquímico, Yago Bilu, Mandinha Damasceno,
Fernandinha Cunha, Rafaello, Ju Gomes, Axa, Renatinha Vilar, Gyovanni, Fernandinha Luz,
Karst Carneiro, Dorien, Erick Oblitas, Sheilita, Nathan, Iokanam, Ana Cláudia, Cláudio e Jéssica
Campos, que sempre com vocês as festas eram muito boas e animadas, com várias recordações
hilárias. Aos pós docs Lucia Fuchsluger, Katrin Fleischer, Sassá Garcia e Bruno Takeshi pela
ajuda com muita sabedoria na ciência. Ao little boss Fumaça, que sempre se dispôs a nos ajudar,
em duvidas do R, nos campos, rancho sucesso no 41, e assegurar que tudo desse certo sempre,
muito obrigada.
Aos pesquisadores José Luís Camargo, João Vitor Rodrigues, Newton Leal, Fabrício
Baccaro, Tânia Pimentel, Ian Hartley, Ben Turner, Patrick Meir, Tomás Domingues, Lazlo Nagy,
Bruce Nelson. Muito obrigada, por doarem seus tempos para me ajudar com novas ideias,
direções, dicas, correções da dissertação e por participarem da minha banca.
Aos parceiros de campo e guerreiros de selva, Luciano, Serjão, muito companheiros,
sempre de bom humor, tudo sempre “bão demais”, sempre tranquilos. Para os queridos Arlindo,
Dona Antoniêta pelas comidas maravilhosas, espetaculares, sem condições de descrever. Ao
Alexandre, Francisco, Pedrinho, Juruna sempre com muita força, apoio e sorriso no rosto. Sem
vocês os campos seriam muito mais difíceis e penosos, seus conhecimentos de vida e de como
sobreviver na floresta também foram essenciais para os nossos dias sem contato com o mundo
exterior, e que fossem mais animados e menos estressantes. Toda essa experiência me renderam
boas memórias em campo, correndo de formigas assassinas e porcos do mato, analisando estrelas
e avistando ETs.
Aos meus amigos queridos Michael, Mandinha Vieira, Tati, Fer, Pri, Ramayane, Iago,
Dani Nogueira, Ana Flávia, Helender, e familiares, que mesmo sem ter a menor ideia do que eu
faço até hoje, me apoiam incondicionalmente nas aventuras no “matinho”, sempre desejando o
melhor e esperando que eu volte inteira dos campos, e ouvir minhas histórias mais loucas.
E agradecimento ao PDBFF, por disponibilizar o amado acampamento 41 para a AFEX
2.0, com piscina natural, cantos de pássaros nas lindas manhãs e de sapos e guaribas adoráveis no
meio da noite.
A turma CFT/2016 por terem me acompanhado nas matérias do INPA. A CAPES/Fapeam
pela bolsa de estudos. Ao Programa de Pós-Graduação em Ciências de Florestas Tropicais do
Instituto Nacional de Pesquisas da Amazônia (CFT/INPA) e todo o seu corpo docente e técnico.
11
Resumo Título: Herbivoria e produção de serapilheira em resposta a adição de nutrientes em floresta de
terra firme não alagada na Amazônia Central.
O papel dos herbívoros na determinação da produtividade das florestas tropicais,
balanço de carbono e ciclagem de nutrientes é incerto na Amazônia. O consumo de folhas
por herbívoros e sua consequente excreção podem representar um fluxo adicional da
serapilheira, muitas vezes não documentado, de nutrientes e carbono no ecossistema. O
foco desta pesquisa foi compreender o papel das limitações nutricionais sobre a
produtividade do dossel no nível do ecossistema, levando em consideração as perdas por
herbivoria e investimento de folhas em compostos secundários em uma floresta de terra
firme na Amazônia Central. O local do estudo foi no projeto AFEX/ PDBFF, um
experimento em larga escala que visa compreender as respostas do ecossistema à
limitação de nutrientes no solo. Os tratamentos consistiram na adição de nitrogênio (N),
fósforo (P), cátions, N + P, N + cátions, P + cátions, N + P + cátions e controle. Os oito
tratamentos foram distribuídos em quatro blocos, representando o desenho fatorial
completo (n = 32 parcelas). A serapilheira foi coletada quinzenalmente por oito meses
(julho/ 2017 e fevereiro/ 2018), utilizando cinco coletores de serapilheira suspensas a um
metro de altura, com área de 0,25 m² cada, totalizando 160 coletores, distribuídos na área
central da parcela de 30 x 30 m. Biomassa de serapilheira, área foliar e perda de área
foliar, macro e micronutrientes, lignina, celulose e polifenóis foram determinados por
análises químicas. Encontramos uma tendência para o aumento da produção de biomassa
de serapilheira no tratamento com N + P durante a estação seca de julho a agosto/ 2017,
e obteve maior input de nutrientes retornando aos solos. Houve um forte aumento nas
concentrações de fósforo, magnésio, cálcio, potássio e manganês nas folhas de
serapilheira em parcelas onde o fósforo foi adicionado. Em contraste com as fortes
mudanças nas concentrações de nutrientes com a fertilização, não houve diferenças
significativas na concentração de compostos de defesa. Como resultado, não houve
diferenças significativas na porcentagem de área foliar perdida para herbívoros (variando
de 9 a 10%) entre os tratamentos. Também não houve diferença significativa na
quantidade de biomassa de herbivoria, porém houveram diferenças significativas de
nutrientes de fezes de insetos, nos tratamentos com adição de P. Juntas, as respostas
iniciais da produção de serapilheira fina sugerem mudanças rápidas nos inputs de
nutrientes da serapilheira, com respostas mais lentas das defensas anti-herbivoria
juntamente com herbivoria às adições de nutrientes.
Palavras-chave: Amazônia, serapilheira, ciclagem de nutrientes, herbivoria,
fertilização.
12
Abstract Title: Herbivory and litter production in response to nutrient addition in a terra
firme forest in Central Amazonia.
The role of herbivores in determining tropical forest productivity, carbon balance
and nutrient cycling is uncertain in the Amazon. Leaf consumption by herbivores and their
consequent excretion may represent an additional, but often undocumented, flow of
nutrients and carbon into the ecosystem. The focus of this research was to understand the
role of nutritional limitations on canopy productivity at the ecosystem level, taking into
account losses by herbivory and leaf investment in secondary compounds in a “terra
firme” forest in Central Amazonia. The study site was the AFEX project, a large scale
experiment that aims to understand ecosystem responses to soil nutrient limitation. The
treatments consisted in the addition of nitrogen (N), phosphorus (P), cations, N + P, N +
Cations, P + Cations, N + P + Cations and control plots. The eight treatments were
distributed in four blocks, where each block contains eight plots representing the complete
factorial design (n= 32 plots). Litter was sampled biweekly for eight months (between
July/ 2017 and February/ 2018), using five suspended littertraps at one meter height, with
an area of 0.25m² each, for a total of 160 collectors. Littertraps were distributed in the
central 30 x 30 m area of each plot. Litterfall biomass, leaf area and leaf area loss, macro
and micronutrient and lignin, cellulose and polyphenols were determined. We found a
trend for increased production of leaf litter biomass in the N + P treatment during the dry
season from July to August/2017 (dry season). This increase in N+P, translated through
litterfall with more nutrients inputs returning to soils. There was a strong increase in
phosphorus, magnesium, calcium, potassium and manganese concentrations in the litter
leaves in plots where phosphorus was added. In contrast to the strong changes in nutrient
concentrations with fertilization, there were no significant differences in the concentration
of defense compounds. Furthermore, there were no significant differences in the percent
leaf area lost to herbivores (varying from 9 to 10%) among the treatments. There was also
no significant difference in the biomass of herbivore frass. However, there were
significant differences of nutrients in the frass, mostly in treatments with P addition.
Together, the intial responses of fine litterfall production suggest rapid changes in litter
nutrient inputs, but slower responses of herbivores and anti-herbivore defenses to nutrient
additions.
Key-words: aboveground productivity, Amazon, herbivory, fertilization,
largescale experiment, litterfall, nutrient cycling, nutrient limitation, tropical forest.
13
1. Introdução
As florestas tropicais desempenham um papel fundamental no ciclo global do
carbono (C) (Cramer et al., 2004), com a própria floresta amazônica respondendo por
cerca de 25% da produtividade primária líquida global. Toda a bacia amazônica cobre 600
milhões de hectares, com 390 bilhões de árvores, que armazenam cerca de 120 petagramas
de C (Pan et al., 2013; Fauset et al., 2015; Zhao e Running, 2011). Os solos ao longo da
bacia amazônica variam muito em geologia e disponibilidade de nutrientes, com um claro
gradiente de fertilidade aumentando de leste para oeste, esta variação pode afetar a
produtividade florestal (Quesada et al., 2010, 2011, 2012).
A idade do solo e as condições climáticas quentes e úmidas da bacia Amazônica
Central contribuem para a transformação do material de origem, gerando solos altamente
intemperizados. A concentração de nutrientes derivados da rocha, como fósforo (P),
cálcio (Ca), magnésio (Mg) e potássio (K), tendem a diminuir, tornando-se
gradativamente menos disponíveis e lixiviados, limitando sua disponibilidade às plantas
(Quesada et al. al., 2010; 2011). Em contraste, as entradas de nitrogênio (N) no solo
ocorrem através da fixação atmosférica de N2 pelos microrganismos do solo, resultando
em acúmulo de N ao longo do tempo. Assim, em solos de florestas tropicais antigas, o N
é encontrado em grandes quantidades em comparação com outros nutrientes do solo,
sugerindo que as florestas tropicais da Amazônia não são limitadas pelo N (Quesada e
Lloyd, 2016; Lambers et al., 2008) .
Em solos amazônicos de baixa fertilidade, a conservação eficiente e o ciclo de
nutrientes é importante para a manutenção do funcionamento da floresta, que é em grande
parte devido à decomposição da serapilheira (Vitousek , 1984). A produção e a rápida
decomposição da serapilheira são processos críticos para o ciclo de nutrientes e a
transferência de energia entre plantas e solo (Quesada et al., 2010; Sayer et al., 2012). Um
ciclo de nutrientes rápido e eficiente também ajuda a reduzir a lixiviação, impedindo que
a floresta em solos de baixa fertilidade se tornem ainda mais pobres em nutrientes (Went
e Stark, 1968; Luizão , 1989).
A serapilheira é definida como a camada de matéria orgânica encontrada em
diferentes estágios de decomposição, formada por folhas, flores, frutos e pequenos ramos
(DAP
14
A produção de serapilheira varia durante o ano com padrões sazonais de
precipitação e é afetada por eventos de seca e chuvas (Wu et al., 2016). Segundo Wu
(2016), a produção de serapilheira é baixa na Amazônia central durante a estação chuvosa
(novembro a maio) e a produção de serapilheira principalmente nos meses secos (agosto
e setembro) tem alta liberação de nova folha. O aumento da irradiância durante o período
de baixa precipitação coincide com o pico de produção de novas folhas para diversas
espécies, devido à resposta da vegetação ao estresse hídrico, uma vez que a perda de folhas
reduz a perda de água pela transpiração (Martins e Rodrigues, 1999). Durante a estação
seca, a capacidade fotossintética do dossel e a produtividade primária também
aumentaram com a expansão da área foliar (Restrepo- Coupe et al., 2013 , Albert et al.,
2018 ). No final da estação chuvosa, o declínio da taxa fotossintética ocorre devido à
senescência foliar, onde os nutrientes são reabsorvidos ou remobilizados antes da abcisão
das folhas, uma importante estratégia nos solos inférteis para a conservação de nutrientes
( Huete et al., 2006 ; Vergutz et al. 2012 ).
A produtividade florestal é influenciada por fatores como a luz, a fenologia foliar,
a disponibilidade de água e nutrientes (Restrepo- Coupe et al., 2013, Borchert et al., 2015).
Para executar as várias funções, as plantas precisam de um conjunto de diferentes
nutrientes. O nitrogênio é essencial no metabolismo vegetal, regulando a fotossíntese, a
assimilação de carbono na enzima rubisco, a manutenção celular, um componente
importante dos ácidos nucléicos, aminoácidos e proteínas, e alguns metabólitos
secundários, bem como a degradação de substâncias ( Taiz e Zeiger , 2009; Marenco e
Lopes, 2011; Lawlor, 2002). Os cátions funcionam como ativadores enzimáticos nas
plantas e também fazem parte da composição das membranas celulares. Mais
especificamente, o cálcio participa da síntese dos tecidos da parede celular e auxilia nas
divisões celulares ( Taiz e Zeiger , 2009). O magnésio desempenha um papel importante
na ativação de enzimas envolvidas na respiração, fotossíntese e síntese de DNA e RNA,
sendo também parte da estrutura da clorofila ( Taiz e Zeiger , 2009). O potássio é
responsável pelo movimento eletroquímico e pela abertura das células estomáticas (
Nunes et al., 2013).
Além disso, uma variedade de compostos orgânicos também são produzidos pelas
plantas, chamados metabólitos secundários (terpenos e fenóis), e sua produção é
influenciada pelos recursos disponíveis (Coley, 1985). Esses metabólitos, também
conhecidos como compostos de defesa, podem interagir nas funções reprodutivas das
plantas, atraindo polinizadores, afetando a competição planta-planta e atuando como
defensivos químicos contra herbívoros e defesas estruturais diminuindo a palatabilidade.
Esses dois tipos de compostos de defesa ocorrem na planta. variando de acordo com o
15
ambiente e idade da folha da planta (Coley et al., 1985; Endara e Coley, 2011 ; Zangerl et
al., 2002 ).
Compostos de defesa da planta podem ser classificados como móvel e não-móvel
e a sua abundância em folhas varia com a idade da planta e renovação do tecido (Coley et
ai., 1985). As defesas móveis, como alcalóides, taninos e fenóis, exigem alto
investimentometabólico , mas têm a vantagem de serem remobilizadas antes da
senescência foliar (Chapin, 1980). Lignina e celulose são compostos estruturais de plantas
que agem para aumentar a dureza foliar com a maturação, reduzindo a palatabilidade
foliar de insetos herbívoros, mas estes são compostos não móveis e não podem ser
remobilizados dentro da planta (Coley et al., 1985 ; Zangerl et al. al., 2002 ).
Outra característica importante da função foliar no crescimento das plantas e um
importante indicador das estratégias de defesa das plantas é a massa foliar por unidade de
área (LMA) ( Lambers e Poorter , 1992; Grime, 2001; Westoby et al., 2002). O LMA está
correlacionada com características fisiológicas e influenciada por recursos ambientais no
campo (Wright et al., 2004). Quando os componentes das características da folha ( por
exemplo, compostos de defesa) estão em maior quantidade por alguma razão ambiental,
a concentração por unidade seria duas vezes maior para todos os constituintes, dobrando
a LMA ( Poorter e Bergkotte , 1992; Van Arendonk e Poorter , 1994). Juntamente com as
defesas secundárias, folhas com alta LMA parecem ter uma melhor defesa contra
herbívoros e perigos físicos (Onoda et al., 2017). Mesmo pequenas concentrações de
defesas secundárias que dificilmente afetam o LMA podem restringir significativamente
o conjunto herbívoro capaz de se alimentar de certas folhas (Coley, 1983).
Herbivoria é o consumo de tecidos vegetais por animais. Existem diferentes tipos
de herbivoria que podem variar entre mamíferos e insetos. Mamíferos maiores (como
macacos e preguiças) representam 25% dos vertebrados herbívoros, e seu consumo é
difícil de medir porque eles removem completamente ramos e folhas (Metcalfe et al.,
2014). Os insetos, com cerca de 500.000 espécies, são considerados os principais
herbívoros, pois representam cerca de 75% da herbivoria (Metcalfe et al., 2014, Herrera
e Pellmyr , 2002, Thomanzini eThomanzini 2000, Novontny e Missa 2000). Na maioria
dos ecossistemas, os tipos mais comuns de herbivoria podem ser identificados como
sugadores foliares ou consumidores de tecido foliar (Hochuli, 2001). A herbivoria é maior
em folhas jovens que crescem durante a estação seca, sendo o principal substrato para
herbívoros (Coley e Barone 1996, Wu et al., 2016). O dano total na área foliar é de cerca
de 12% na floresta tropical, com cerca de 0,0003-0,8% da superfície foliar sendo
consumida por dia, dependendo da espécie arbórea (Coley, 1983; Coley e Barone, 1996).
Em florestas montanhosas ao longo de um transecto entre a Amazônia e os Andes, a
herbivoria tem mostrado afetar de 12 a 19% da produtividade das folhas (Metcalfe et al.,
16
2014). O principal fator que controlou a variação nas taxas de herbivoria com elevação
(200 a 3400 manm .) Foi a diferença na temperatura e na concentração foliar de nutrientes
entre as florestas (Metcalfe et al., 2014).
As plantas também podem adotar mecanismos para evitar perdas de nutrientes da
herbivoria, tais como rápida expansão da folha e maturação do cloroplasto durante a
expansão, produção de folhas síncronas, compostos de defesa e várias formas de
mutualismo e antagonismo (Coley e Barone, 1996). Um estudo de Werner e Homeier
(2015) em uma floresta tropical de altitude mostrou que a quantidade de N e P nas folhas,
pode influenciar a área foliar removida pela herbivoria. Além disso, é importante observar
que o manejo de nutrientes e gradientes ambientais indicam que as concentrações de
nutrientes do solo em florestas tropicais podem influenciar os níveis de nutrientes nos
tecidos vegetais (Sayers et al., 2012, Sullivan et al., 2014; Coley e Kursar , 1996; Throop
e Lerdau , 2004). Além disso, o experimento de fertilização no Panamá mostrou que as
taxas de herbivoria diferiam dependendo do nutriente que foi adicionado (Santiago et al.,
2012). Mais herbivoria foi observada com adição de P e K, quando comparado ao N,
sugerindo que o N não foi um nutriente limitante para herbívoros neste local.
Muitos herbívoros preferem novos tecidos foliares por reduzir as concentrações de
dureza folha e nutrientes de N, P, Mg, K, a água, o conteúdo de hidrato de carbono s, em
comparação com as folhas maduras, e mudanças na fenologia folha pode afetar as
interacções de insectos de plantas (Coley, 1980; 2005). Quantidade de grau superior H de
polifenóis na folha está relacionada a menores taxas de herbivoria ( Markow et al., 1999).
A concentração de N na folha pode servir como um bom preditor de herbivoria, uma vez
que existe uma forte relação positiva entre a concentração de N e o ciclo de vida dos
insetos na planta. e Lerdau , 2004. Além de N que aumenta a palatabilidade das folhas
para insetos, outros nutrientes também são importantes, como P, K e Mg, que são
abundantes em tecidos de insetos (Mattson e Scriber, 1987; Clancy e King, 1993).
Os herbívoros de insetos também podem afetar o ciclo de nutrientes no solo com
deposição de fezes de inseto nos insumos de superfície do solo (Hunter, 2001, Riesley e
Crossley, 1988, Hunter et al., 2003). Alguns estudos que descrevem a importância
deexcrementos de insectos no ciclo dos nutrientes (Bardgett e Wardle, 2003) mostram que
a deposição dos excrementos resultou num aumento da mineralização N (Lightfoot e
Whitford, 1990; Reynolds et ai, 2000), o aumento da imobilização microbiana de N em
microcosmos e experimentos de campo (Lovett e Ruesink , 1995).
Houve apenas quatro experimentos em grande escala que tentam entender o efeito
da limitação nutricional no ciclo do carbono em florestas tropicais de florestas maduras:
Bornéu ( Mirmanto et al. 1999), Camarões (Newberry et al 2002), Panamá (Wright et al.
al. 2011) e Costa Rica (Alvarez-Clare et al., 2013). Mesmo com as adições de nutrientes,
as respostas de crescimento do tronco de árvore são relativamente baixas ( Mirmanto et
17
al., 1999, Wright et al., 2011, 2018, Alvarez-Clare et al 2013), uma possível explicação
para respostas fracas de crescimento às adições de nutrientes é aumentada pressão de
pragas nas parcelas fertilizadas (Campo e Dirzo 2003, Andersen et al., 2010, Santiago et
al., 2012, Wright et al., 2018).
Um estudo de Werner e Homeier (2015) em uma floresta tropical de altitude
mostrou que a quantidade de N e P nas folhas, pode influenciar a área foliar removida
pela herbivoria. Além disso, é importante observar que o manejo de nutrientes e gradientes
ambientais indicam que as concentrações de nutrientes do solo em florestas tropicais
podem influenciar os níveis de nutrientes nos tecidos vegetais (Sayers et al., 2012,
Sullivan et al., 2014; Coley e Kursar , 1996; Throop e Lerdau , 2004). Além disso, o
experimento de fertilização no Panamá mostrou que as taxas de herbivoria diferiam
dependendo do nutriente que foi adicionado (Santiago et al., 2012). Mais herbivoria foi
observada com adição de P e K, quando comparado ao N, sugerindo que o N não foi um
nutriente limitante para herbívoros neste local.
A fertilização geralmente aumenta o conteúdo de nutrientes dos tecidos vegetais
(Sayer et al., 2012), potencialmente tornando as folhas suscetíveis à herbivoria. De fato,
vários estudos mostraram aumento nas taxas de herbivoria com o aumento de nutrientes
foliares para plântulas crescendo em parcelas fertilizadas (Andersen et al., 2010, Santiago
et al., 2012). No entanto, não se sabe se o aumento herbivoria com adição de nutrientes
pode afetar um dossel nas taxas de crescimento de árvores ou como efeitos herbivoria
adaptar-se ao nível do ecossistema (Metcalfe et al., 2014). Aqui, tentamos responder à
questão de saber se a herbivoria no nível do ecossistema muda com a adição de nutrientes
e, portanto, possivelmente medeia o crescimento das árvores, concentrando-se na
produção de dossel e na herbivoria.
Este estudo examinou a limitação de nutrientes em um local de baixa fertilidade
do solo na Amazônia Central, quantificando a função florestal e a produtividade em
resposta à fertilização. N, P, cátions (Ca, Mg e K) e o experimento de fertilização da
Amazônia (AFEX)foram instalados em uma floresta madura em um dos tipos de solo
mais difundido em toda a Amazônia e são o primeiro experimento em grande escala. Dada
a baixa disponibilidade de P e cátions na Amazônia Central, este experimento está
preparado para melhorar nossa compreensão da limitação de nutrientes em toda a
Amazônia.
18
1.1. Introduction
Tropical forests play a key role in the global carbon cycle (C) (Cramer et al., 2004), with
the Amazon rainforest itself accounting for about 25% of global net primary productivity. The
entire Amazon basin covers 600 million hectares with 390 billion trees, which stores about 120
petagrams of C (Pan et al., 2013, Fauset et al., 2015, Zhao and Running, 2011). Soils along the
Amazon basin vary widely in geology and nutrient availability, with a clear fertility gradient
increasing from east to west, this variation could affect forest productivity (Quesada et al., 2010,
2011, 2012).
Soil age and hot and humid climatic conditions of the Central Amazon basin contribute to
the transformation of the source material, generating highly weathered soils. The concentration of
nutrients derived from the rock, such as phosphorus (P), calcium (Ca), magnesium (Mg) and
potassium (K), tend to decrease, becoming gradually less available and leached, limiting their
availability to plants (Quesada et al., 2010; 2011). In contrast, nitrogen (N) inputs in the soil occur
through the atmospheric fixation of N2 by soil microorganisms, resulting in accumulation of N
over time. Thus, in ancient tropical forest soils, N is found in large amounts compared to other
soil nutrients, suggesting that Amazonian rainforests are not limited by N (Quesada and Lloyd,
2016; Lambers et al., 2008 ).
In low fertility Amazonian soils, efficient conservation and cycling of nutrients is
important for the maintenance of forest functioning, which is largely due to the decomposition of
the litter (Vitousek, 1984). The production and rapid decomposition of litter are critical processes
for nutrient cycling and energy transfer between plants and soil (Quesada et al., 2010; Sayer et al.,
2012). A fast and efficient nutrient cycle also helps reduce leaching, preventing forest in low
fertility soils from becoming even poorer in nutrients (Went and Stark, 1968; Luizão, 1989).
The litter is defined as the layer of organic matter found in different stages of
decomposition, formed by leaves, flowers, fruits and small branches (DAP
19
water loss through transpiration (Martins and Rodrigues , 1999). During the dry season, the
photosynthetic capacity of the canopy and primary productivity also increased with leaf area
expansion (Restrepo-Coupe et al., 2013, Albert et al., 2018). At the end of the rainy season, the
decline in the photosynthetic rate occurs due to leaf senescence, where nutrients are reabsorbed
or remobilized before leaf excision, an important strategy in infertile soils for nutrient
conservation (Huete et al., 2006; Vergutz et al. 2012).
Forest productivity is influenced by factors such as light, foliar phenology, water
availability and nutrients (Restrepo-Coupe et al., 2013, Borchert et al., 2015). To perform the
various functions, plants need a set of different nutrients. Nitrogen is essential in plant metabolism,
regulating photosynthesis, carbon assimilation in the rubisco enzyme, cell maintenance, an
important component of nucleic acids, amino acids and proteins, and some secondary metabolites,
as well as degradation of substances (Taiz and Zeiger, 2009; Marenco and Lopes, 2011; Lawlor,
2002). Cations function as enzyme activators in plants are also part of the composition of cell
membranes. More specifically, calcium participates in the synthesis of cellular wall tissues and
assists in cell divisions (Taiz and Zeiger, 2009). Magnesium plays an important role in the
activation of enzymes involved in respiration, photosynthesis and synthesis of DNA and RNA,
being also part of the structure of chlorophyll (Taiz and Zeiger, 2009). Potassium is responsible
for the electrochemical movement and stomatal cell opening (Nunes et al., 2013).
In addition, a variety of organic compounds are also produced by plants, called secondary
metabolites (terpenes and phenols), and their production is influenced by available resources
(Coley, 1985). Such metabolites, also known as defense compounds, could interact in the
reproductive functions of plants, attracting pollinators, affecting plant-plant competition and
acting as chemical defenses against herbivores, and structural defenses decreasing palatability
These two types of defense compounds occur in the plant, varying according to the environment
and age of the plant leaf (Coley et al., 1985; Endara and Coley, 2011; Zangerl et al., 2002).
Plant defense compounds can be classified as mobile and non-mobile and their abundance
in leaves varies with plant age and tissue turnover (Coley et al., 1985). Mobile defenses, such as
alkaloids, tannins and phenols, require high metabolic investment, but have the advantage of being
remobilized prior to foliar senescence (Chapin, 1980). Lignin and cellulose are structural
compounds of plants that act to increase foliar toughness with maturation, reducing the foliar
palatability of herbivorous insects, but these are non - mobile compounds and can not be
remobilized within the plant (Coley et al., 1985; Zangerl et al., 2002).
Another important feature of leaf function on plant growth and an important indicator of
plant defense strategies is leaf mass per unit area (LMA) (Lambers and Poorter, 1992; Grime,
2001; Westoby et al., 2002). LMA is correlated with physiological characteristics and influenced
by environmental resources in the field (Wright et al., 2004). When components of leaf
20
characteristics (eg defense compounds) are in greater quantity for some environmental reason, the
concentration per unit would be twice as large for all constituents, doubling the LMA (Poorter
and Bergkotte, 1992; Van Arendonk and Poorter , 1994). Together with the secondary defenses,
leaves with high LMA seem to have a better defense against herbivores and physical hazards
(Onoda et al., 2017). Even small concentrations of secondary defenses that hardly affect LMA
may significantly restrict the herbivorous set capable of feeding on certain leaves (Coley, 1983).
Herbivory is the consumption of plant tissues by animals. There are different types of
herbivory that can vary between mammals and insects. Larger mammals (such as monkeys and
sloths) represent 25% of herbivorous vertebrates, and their consumption is difficult to measure
because they completely remove branches and leaves (Metcalfe et al., 2014). The insects, with
about 500,000 species, are considered the main herbivores because they account for about 75%
of the herbivory (Metcalfe et al., 2014, Herrera and Pellmyr, 2002, Thomanzini and Thomanzini
2000, Novontny and Missa 2000). In most ecosystems, the most common types of herbivory can
be identified as suckers and leaf strippers or consumers of leaf tissue (Hochuli, 2001). The
herbivory is larger in young leaves growing during the dry season, being the main substrate for
herbivores (Coley and Barone 1996, Wu et al., 2016). The total leaf area damage is about 12% in
the rainforest, with about 0.0003-0.8% of the leaf surface being consumed per day, depending on
the tree species (Coley, 1983; Coley and Barone, 1996 ). In mountainous forests along an
Amazon-Andes transect, herbivory has been shown to affect 12 to 19% of leaf productivity
(Metcalfe et al., 2014). The main factor that controlled the variation in herbivory rates with
elevation (200 to 3400 m.a.n.m.) was the difference in temperature and foliar P concentration of
nutrients among forests (Metcalfe et al., 2014).
Plants can also adopt mechanisms to avoid nutrient losses from herbivory, such as rapid
leaf expansion and chloroplast maturation during expansion, synchronous leaf production, defense
compounds and various forms of mutualism and antagonism (Coley and Barone, 1996). A study
by Werner and Homeier (2015) in a tropical montane rainforest showed that the amount of N and
P in leaves, can influence the leaf area removed by herbivory. In addition, it is important to note
that nutrient handling and environmental gradients indicate that soil nutrient concentrations in
tropical forests can influence the levels of nutrients in plant tissues (Sayers et al., 2012, Sullivan
et al., 2014; Coley and Kursar, 1996; Throop and Lerdau, 2004). Additionally, the fertilization
experiment in Panama showed that herbivory rates differed depending on the nutrient that was
added (Santiago et al., 2012). More herbivory was observed with P and K addition, when
compared to N, suggesting that N was not a limiting nutrient for herbivores at this site.
Many herbivores prefer new leaf tissues due to lower leaf toughness and nutrients
concentrations of N, P, Mg, K, water, carbohydrates content, compared with mature leaves, and
changes in leaf phenology may affect insect-plant interactions (Coley, 1980; 2005). Higher
amount of polyphenols in the leaf is related to lower herbivory rates ((Markow et al., 1999). Leaf
21
N concentration can serve as a good predictor of herbivory, since there is a strong positive
relationship between the N concentration and the life cycle of insects in the plant (Throop and
Lerdau, 2004). In addition to N which increases leaf palatability for insects, other nutrients are
also important, such as P, K and Mg which are abundant in insect tissues (Mattson and Scriber,
1987; Clancy and King, 1993).
Insect herbivores can also affect soil nutrient cycling with frass (insect feces) deposition
in surface soil inputs (Hunter, 2001, Riesley and Crossley, 1988, Hunter et al., 2003). A few
studies that discuss the importance of insect frass in nutrient cycling (Bardgett and Wardle, 2003)
show that frass deposition has resulted in increased N mineralization (Lightfoot and Whitford,
1990; Reynolds et al., 2000), increased microbial immobilization of N in microcosm and field
experiments (Lovett and Ruesink, 1995).
There have only been four large-scale experimental tests that attempt to understand the
effect of nutritional limitation on the carbon cycle in mature lowland tropical forests: Borneo
(Mirmanto et al. 1999), Cameroon (Newberry et al 2002), Panama (Wright et al. 2011), and Costa
Rica (Alvarez-Clare et al., 2013). Even with the nutrient additions the tree trunk growth responses
are relatively weak (Mirmanto et al., 1999, Wright et al., 2011, 2018, Alvarez-Clare et al 2013),
a possible explanation for weak growth responses to nutrient additions is increased pest pressure
in the fertilized plots (Campo and Dirzo 2003, Andersen et al., 2010, Santiago et al., 2012, Wright
et al., 2018).
A study by Werner and Homeier (2015) in a tropical montane rainforest showed that the
amount of N and P in leaves, can influence the leaf area removed by herbivory. In addition, it is
important to note that nutrient handling and environmental gradients indicate that soil nutrient
concentrations in tropical forests can influence the levels of nutrients in plant tissues (Sayers et
al., 2012, Sullivan et al., 2014; Coley and Kursar, 1996; Throop and Lerdau, 2004). Additionally,
the fertilization experiment in Panama showed that herbivory rates differed depending on the
nutrient that was added (Santiago et al., 2012). More herbivory was observed with P and K
addition, when compared to N, suggesting that N was not a limiting nutrient for herbivores at this
site
Fertilization generally increases nutrient content of plant tissues (Sayer et al., 2012),
potentially making leaves susceptible to herbivory. Indeed, several studies have shown increased
herbivory rates with increased leaf nutrients for seedlings growing in fertilized plots (Andersen et
al., 2010, Santiago et al., 2012). However, it is uncertain whether the increased herbivory with
nutrient addition can affect canopy tree growth rates or how herbivory effects scale up to the
ecosystem level (Metcalfe et al., 2014). Here, we attempt to answer the question of whether
ecosystem-level herbivory changes with nutrient addition, and therefore possibly mediates tree
growth, by focusing on canopy production and herbivory.
22
This study examined nutrient limitation in a low soil fertility site in Central Amazonia by
quantifying forest function and productivity in response to fertilization. N, P, Cations (Ca, Mg
and K) and the Amazonian Fertilization Experiment (AFEX) were installed in a mature forest on
one of the most widespread soil types across Amazonia and are the first large-scale experiment.
Given the low availability of P and cations in Central Amazonia, this experiment is poised to
enhance our understanding of nutrient limitation across Amazonia.
2. Objective 2.1. General Objective
Understand the role of nutritional limitation on canopy productivity, herbivory loss
and investment in secondary compounds in terra firme forests in Central Amazonia, at the
ecosystem scale.
2.2. Specific Objectives
1- Determine the response of fine litterfall productivity to nutrient addition;
2- Determine the response of leaf traits, including LMA, the concentration of
macro and micronutrients as well as secondary compounds to fertilization;
3- Determine the effect of nutrient addition on leaf herbivory at the ecosystem
level and the feedback of herbivore nutritional contribution across the different
treatments;
4- Determine the relationships between litter production, investment in
secondary compounds, concentration of nutrients and levels of herbivory.
3. Hypotheses
H1: If nutrient additions increase leaf quality, then leaf turnover expected to
increase which will result in higher leaf litter productivity in fertilized compared
to control treatments;
H2: If nutrient additions increase leaf quality, then leaf litter nutrient content will
increase, whereas LMA and the investment in defense compounds are expected
to decrease, resulting in greater palatability, and decreased leaf toughness;
H3: If nutrient additions increase leaf quality, and leaf quality determines both
litter production and herbivory levels, a positive correlation between litter
production and herbivory levels is expected.
23
4. Methods
4.1. Location and characterization of the study area
The study took place in the ZF-3 Reserve area called "Km 41", within the AFEX
project area (Figure 1) at the Biological Dynamics of Forest Fragments Project (BDFFP/
INPA) located approximately 80 km north of Manaus/Amazonas/Brazil (02° 25' 00'' S;
59° 43' 00' W). The climate in the region is characterized by small seasonal variation in
air temperature, with an average of 26.7 °C. Mean annual rainfall ranges from 1900 mm
to 3500 mm, with a peak in April (> 300 mm/month) and a dry period (
24
ombrophilous forest (Laurance et al. 2018). The average height of the canopy varies from
30-37 meters with occasional emergent trees that can reach 55 meters (Oliveira and Mori,
1999). The area has an estimated minimum of 280 tree species ha-1 distributed in 53
families. Plant families with higher density in the area are Lecythidaceae, Fabaceae,
Sapotaceae and Burseraceae. (Rankin–de-Merona et al., 1992).
AFEX project started in May 2017 as a full factorial experiment with the additions
of Nitrogen (N), Phosphorus (P) and Cations and their combinations. The experiment has
eight treatments (TABLE 1) and four replicates per treatment, resulting in a total of four
(4) independent blocks (at least 250 m distance between each other) and thirty-two (32)
plots, 50 m apart from each other (FIGURE 2). These treatments allow the analysis of
potential limitations of the different essential elements on plant growth (canopy
productivity), nutrient content of plant tissues and herbivory. Nutrients were added to the
plots manually, with dry granules applied to the soil surface three times a year, avoiding
the main dry season. Each plot measures 50 m x 50 m where nutrients are applied, with
the main measurements limited to the central plot area (30 m x 30 m) (FIGURE 3), thus
aiming to maximize the probability that the litter sampled belonged to trees inside the
plots. Nutrients were added at the following rates: 125 kg ha-1 year-1 of N as urea, 50 kg
ha-1 year-1 of P as triple superphosphate, 50 kg ha-1 year-1 of Ca and 20 kg ha-1 year-1 Mg
as dolomitic limestone and 50 kg ha-1 year-1 of K as potassium chloride.
FIGURE 2. Example of the position diagram of plots inside one block with trails between the eight
plots in blue line, and the treatments.
TABLE 1. List of treatments names in the factorial design used in AFEX fertilization.
Treatments used in AFEX fertili zation.
Control Cations
Nitrogen (N) Nitrogen (N) + Cations
Phosphorus (P) Phosphorus (P) + Cations
Nitrogen (N) + Phosphorus (P) Nitrogen (N) + Phosphorus (P) + Cations
25
TABLE 2. Concentrations of the nutrients Nitrogen, Phosphorus and Cations in the treatments Concentrations of nutrients in treatments.
Nitrogen (N) 125 kg ha -1 year -1 , as urea
Phosphorus (P) 50 kg ha -1 year -1 in the form of triple superphosphate
Cations Ca+ 50 e Mg +, 20 kg ha-1 year-1 dolomitic limestone
K+, 50 kg ha -1 year -1 as potassium chloride
For the determination of litter production, litter traps were built using plastic tubes,
with dimensions of 50 cm x 50 cm, occupying an area of 0.25 m2 and one meter above
the ground level. Five litter traps were installed at five (5) locations within the central area
of each plot to insure litter reaching the trap was produced within the experimental plot
area (FIGURE 3). To capture senesced leaves fallen from the canopy, we used
polyethylene screens of 1 mm mesh. The installation of the littertraps occurred in the first
week of July 2017, and the first collection was mid July. Samples were collected biweekly
for eight months, making a total of 13 census collections ending in February 2018.
FIGURE 3.Picture of the plot design of 50 m² in grey, with the buffer zone of 10 m in red dotted lines, the central area inside the plot with 30m² in green, and distribution of the five litter traps
in yellow.
26
FIGURE 4.a) Littertrap installed in the plot; b) litterfall collection in one of the five littertraps in
the center plot area. Source: Andersen, K and Moraes, A.C.M.
4.3. Calculations
4.3.1. Litterfall biomass
After every collection, fine litter was sorted at INPA and separated into different
fractions: leaves, twigs and thin branches (woody material) with a diameter
27
4.3.3. Calculation of leaf mass and area missing by herbivory
Leaf area loss by herbivore was calculated as the proportion of the total leaf area
without herbivory (AT in m²), the original leaf area collected in the field with herbivory
(real area (AR in m²) generating by difference the proportion of leaf lost by herbivory
(LH). The potential total leaf production in the canopy Mg ha-1 year-1 (Pr) was calculated
by dividing the known production of the litter of the collectors (PH) minus the proportion
of biomass removed by herbivore by 1 – LH. To determine the effect of leaf herbivory on
canopy productivity (Mg ha-1 year-1) we calculated biomass consumed by herbivory Mg
ha-1 year-1 (Bh) (Metcalfe et al. 2014; Werner and Homeier 2015, Cárdenas et al. 2014).
FIGURE 5. Example of leaves with and without herbivory used to calculate leaf area loss. a) is the
scanned leaf original area measured (with herbivory); b) is the calculation of the area missing projected in
red at the program Image J; c) is the leaf designed without herbivory. Source: K. Andersen e Moraes,
A.C.M.
4.4. Chemical analysis
4.4.1. Macro and micro nutrients
To determine macro and micronutrients and secondary compounds, litter samples
were bulked by plot using material collected during the months of August and September
2017. All analysis were conducted at INPA's soil and plants lab. After dried, samples were
ground using a knife and/ or ball mill for nutritional analysis. Leaves and insect frass were
analysed separately. Micronutrients, phosphorus and cations were analysed by
nitroperchloric digestion protocol described by Malavolta et al. (1989). The bases cations
(K, Ca and Mg) and micronutrients were determined by atomic absorption
𝐿𝐻 % = 𝐴𝑇 − 𝐴𝑅
𝐴𝑇 𝑃𝑟 =
𝑃𝐻 1 − 𝐿𝐻
𝐵 ℎ = 𝑃𝑟 − 𝑃𝐻
28
spectrophotometry (AAS, 1100 B, Perkin-Elmer, Ueberlingen, Germany) (Anderson and
Ingram, 1993). The total P concentrations were determined by colorimetry (Anderson and
Ingram, 1993), and quantified by spectrophotometry (UV-120-01, Shimadzu, Kyoto,
Japan). The C and N contents were determined using an automatic C and N analyzer
(VARIO MAX CHN Element Analyzer) (Nelson and Sommers, 1996). Leaves were not
identified by species and leaf characteristics represent means at plot level (Metcalfe et al.,
2014).
4.4.6. Lignin and cellulose
Concentrations of lignin and cellulose in litter were determined by the ADF-
sulfuric method (Van Soest 1963; Rowland and Roberts, 1994). The method consists in
applying an acidic detergent solution (ADF), followed by the cellulose using H2SO4 72%,
with lignin subsequently determined by weight loss on ignition 0 to 50° C in a muffle
furnace.
4.4.7. Polyphenols
Total polyphenols (simple phenols and hydrolysable tannins) were determined
following the method of Folin Denis (Coley, 1983). Extraction consists of homogenizing
1 g of leaves in 35 ml of 85% methanol solution, then the sample was diluted to the volume
of 50 ml and put in a water bath (70°C) for 20 minutes. The concentration of polyphenols
was determined by colorimetry (Anderson and Ingram, 1993), and quantified by
spectrophotometry (UV-120-01, Shimadzu, Kyoto, Japan).
4.4.1. Calculations leaf litterfall nutrient input
Leaf litterfall nutrient biomass (LNB) was calculated by summing leaf biomass
(g/m²) from August and September (LBD). LBD was then converted to kg/ha-1 and
multiplied by nutrient concentration in litter (LNut):
𝐿𝑁𝐵 = [𝐿𝐵𝐷 ∗ 10] ∗ 𝐿𝑁𝑢𝑡
4.4.2. Herbivory frass nutrient biomass input
To answer the fourth objective of determine the effect of the herbivory frass input
in biomass (kg/ ha-1/ year-1) and nutrients on the treatments, we used the following
equation for frass nutrient biomass input (FNB). Insect frass biomass (FBD)] was
converted to kilogram per hectare and multiplyed by quantity of nutrients found in the
frass (FNut).
𝐹𝑁𝐵 = [𝐹𝐵𝐷 ∗ 10] ∗ 𝐹𝑛𝑢𝑡
29
4.5. Data analysis
The statistical program R version 3.4.4 (R core team, 2018) was used to test the
effect of nutrient additions on the response variables studied here. We conducted a series
of linear mixed models using package “lmer4” (Bates et al., 2015), considering block as
a random factor. The fixed factor was used in two ways, first by treatments comparing the
control against the other seven treatments. When the model was significant the
Dunnett’s test was applied to test treatments separately comparing with control, using the
package “Multcomp” (Hothorn et al., 2008). The second approach was using the factorial
experimental design to obtain the main effect of each nutrient and testing the interactions
between the nutrients. Model simplification procedures were used to find the most
parsimonious model based on AIC values. When the fixed effects were significant in the
final model, a post-hoc tukey test was applied using the package “emmeans” (Lenth,
2018), to test the nutrient interaction effects. The model fit was verified to make sure the
analysis met the model assumption, checking the normality, standardized residuals plot
and density of the model residuals.
To determine the importance of leaf mechanisms against herbivory we tested the
level of investment in secondary compounds with the availability of nutrients. More
specifically, we tested the effect of lignin, cellulose, polyphenols, macro and
micronutrients and specific leaf area on leaf biomass consumed by herbivory (Bh). We
tested the effect of treatments on biomass consumed by herbivory (Bh), nutrient
consumed by herbivory (Nh) and productivity of leaf litterfall biomass (PH). To
determine the relationships between litter production, investment in secondary
compounds, concentration of nutrients and levels of herbivory, we conducted correlation
tests using the package “Hmisc” (Harrell, 2018) to obtain the Pearson Correlation
Coefficient (r). If the test showed high correlation, we further explored the relationship
using a Generalized Linear Mixed Model (GLMM), with the herbivore variables as
response variables and the leaf litter variables as the fixed factors and block as the random
factor. Finally, to better understand the multivariate relationships of leaf litter macro and
micronutrients concentrations (P, Ca, Mg, K, Mn g.kg-1), leaf mass area (LMA) and
defense compounds (Lignin, Cellulose and Phenols) on herbivory responses we used a
principal component analysis (PCA).
30
5. Results
5.1. Patterns of litterfall fractions with treatment
To answer the first objective, that is to determine the response of fine litterfall
productivity to nutrient additions, we gathered information of mean values of fine litterfall
biomass fractions, cumulative fine litterfal biomass, and production of fine litterfall over
time. Over 8 months, with 13 field collections of fine litterfall biomass production, the
values between treatments varied from 9.12±6.33 (P+Cations) to 10.73±8.81 (N+P) Mg
ha-1 year-1 (FIGURE 6). It was found a trend for differences in total litterfall between the
treatments with higher production in the “N+P” treatment (F7,24=2.32,P=0.067). The
litterfall production of the different fractions did not differ significantly among the
treatments. Leaf litterfall varied from 5.73±0.46 (N) to 7.59±0.82 (N+P) Mg ha-1 year-1.
Fine woody litterfall ranged from 1.48±0.17 (P+Cations) to 2.19±0.32 (N+P+Cations) Mg
ha-1 year-1. For reproductive material, productivity varied from 0.82±0.11 (P+Cations) to
1.6±0.24 (P) Mg ha-1 year-1. Other residues varied from 0.08±0.01 (N+P+Cations) to
0.16±0.01 (P) Mg ha-1 year-1. The insect frass fraction varied from 0.05±0.006
(N+Cations) to 0.08±0.009 (P+Cations) Mg ha-1 year-1.
FIGURE 6. Mean fractions of fine litterfall Mg ha-1 year-1 (leaves in green color, fine wood in orange,
reproductive material in pink, other residues in blue, and frass in brown), by nutrient addition treatments of
Control (CTL), Nitrogen (N), Phosphorus (P), Cations (Cation), N+P (NP), N+Cations (NCation), P+Cations
(PCation), and N+P+Cations (NPCation).
Fine litterfall biomass production in all treatments showed a seasonal pattern,
expected for Manaus region (FIGURE 8), with peak production occurring in the first week
of September during the dry season. There were no significant differences among
treatments through the sampling time. The “N+P” treatment showed the highest biomass
31
production and a slightly earlier peak in production in the dry season. Litter production
declined from October 2017 until February 2018, during the transition from dry to wet
season.
FIGURE 7. Fine litterfall (Mg ha-1 year-1) with nutrient addition treatments of Control (CTL),
Nitrogen (N), Phosphorus (P), Cations (Cation), N+P (NP), N+Cations (NCation), P+Cations (PCation),
and N+P+Cations (NPCation). The x axis is counting the collection date over time (july/2017 –
february/2018).
FIGURE 8. Cumulative leaf litterfall(g.m²) over time by treatments Control (CTL), Nitrogen (N),
Phosphorus (P), Cations (Cation), N+P (NP), N+Cations (NCation), P+Cations (PCation), and
N+P+Cations (NPCation).
The cumulative leaf litterfall biomass (FIGURE 9) indicated that the treatment
“N+P” had abscised leaves before the other treatments in the second week of August,
making their biomass stand out and perhaps growing apart from the other treatments since
then. In the middle of November the treatments “N+P+Cations” and “N+Cations”
32
appeared to start to slightly increase the amount of litterfall compared to the other
treatments.
5.2. Leaf Mass Area
To answer the second objective to determinate foliar characteristics, like LMA,
macro and micronutrients, and secondary compounds, we examined how , leaf mass area
(g/m²), nutrient concentrations of leaves (g.kg-1), leaf litter nutrient input (g.ha-1), and
defense compounds (%) change with fertilization.
The leaf mass area had significant interactions between Nitrogen and Cations
(F1,28=4.65,P=0.04*). Specifically, the leaves that received both N and Cations fertilizers
(FIGURE 10a) had a higher leaf mass per unit leaf area, indicating the leaves got thicker
(max = +N+Cations: 191.67±23.84 g.m2; min = +N-Cations: 110,33±12.55 g.m²). The
increase in LMA with added N and cations is particularly pronounced in comparasion to
the other treatments (FIGURE 10b).
FIGURE 10. a show Leaf Mass Area (g.m²) with the factorial interaction of Nitrogen and Cations.
Showing that the LMA had a significant response with the presence of both Nitrogen and Cations. Figure
b show all treatments LMA pattern. Dotted line represent the mean LMA value in control plots. P
33
FIGURE 11. figure a is Leaf litter phosphorus concentration (g.kg-1) with and without factorial P
treatments, showing that phosphorus had a significant response in the presence of P treatments. FIGURE
b shows the values of nutriet P with all treatments. Dotted line represent the mean leaf P content in
control plots. P
34
compared to plots that did not receive it (–P; 27.6±1.77 g.ha-1) (F1,26=13.2; P=0.001**).
The nutrient Manganese was higher in the treatments with P (+P; 24.49±2.08 g.ha-1) in
comparison with not receiving P (-P; 18.26±1.47 g.ha-1), and was significant in factorial
P (F1,29=6.0; P=0.02*).
The leaf litter nutrients inputs of Calcium (F7,21=2.74; P=0.03*) and Magnesium
(F7,24=3.89; P=0.005**) had a similar pattern, and are significantly different from control
in the treatments N+P (Ca 1012±123 g.ha-1 and Mg 406±25 g.ha-1) and P (Ca 974±193
g.ha-1 and Mg 407±59.2 g.ha-1), with the highest mean values. Ca inputs increased
significant in the N+P and +P treatments compared to the control. Potassium was the only
nutrient that did not significant change in response of nutrient additions probably
influenced by the high level of variation in K inputs, with the treatment N+P was the
highest mean value (201±31.1 g.ha-1) (FIGURE 12).
I also calculated the leaf nutrient input in kg.ha-1yr-1 (TABLE 4), these results
assume a constant value throught out the year based on dry season inputs. For P, the input
varied from 1.56±0.21 kg.ha-1 (N+Cations) to 2.49±0.04 kg.ha-1 (N+P), showing that the
mean. For Mg, the input varied from 16.6±1.76 kg.ha-1 (Control) to 27.5±1.5 kg.ha-1
(N+P). For Ca, varied from 37±3.94 kg.ha-1 (P+Cations) to 68.3±7.85 kg.ha-1 (N+P). For
K input, varied 9.83±0.57 kg.ha-1 (Control) to 13.5±2.03 kg.ha-1 (N+P).
FIGURE 10. Amount of phosphorus (g.ha-1) (figure a and b), phosphorus response in the presence
of treatments P, and without the presence of Cations. Manganese (g.ha-1) (figure c) response with factorial
P treatment. P
35
(the mean value in marked with black dotted line), Nitrogen (N) in light green, N+Cations in dark green,
N+P in light blue, N+P+Cations in turquoise blue, Phosphorus (P) in lilac, P+Cations in pink.
P
36
Cations; 0.18±0.02 g.kg-1). However, when scaling to hectares, this value is no longer
significant (F7,21=1.49, P=0.22). The treatment N+Cations (1.6±0.17 g.ha-1), had the
highest mean herbivory value, with the lowest being treatment P+Cations. Together, these
results show that N+Cations was the treatment with more significance in all variables
related to herbivory, probably because their higher mass and area originate from the same
samples.
FIGURE 13. Leaf area missing (%) (a), Leaf Biomass Herbivory (kg.ha-1) (b) in all 8 treatments
in the x axis. Showing the control mean value in marked with black dotted line. P
37
5.8. Insect Frass Nutrient content
Analyzing the nutrients of the insect frass, the nutrients that had significance using
the factorial model were phosphorus (F1,26=4.49; P=0.04*), and manganese (F1,28=7.8;
P=0.009**). For phosphorus in insect frass, the treatments that did not receive added
Cations (-Cations; 0.63±0.035 g.kg-1) were higher than those that received added Cations
(+Cations; 0.54±0.06 g.kg-1) this is where you include the LMM result.
FIGURE 15. The Insect Frass Nutrient (g.kg-1), showing in figure “a” the nutrient phosphorus,
figure “b” the nutrient calcium, figure “c” the nutrient potassium, figure “d” magnesium and “e”
manganese.. The x axis being all treatments, the mean control value in marked with black line. P
38
nutrients inputs did not differ among treatments for potassium (F7,24=0.61; P=0.73)
(FIGURE 18).
FIGURE 16. Quantity of P (a),Ca (b) , Mg (d), Mn (e) and K (c) in insect biomass frass (g.ha-1)
with and without P, and in all 8 treatments. The x axis being all treatments and the mean control value in
marked with black line. P
39
FIGURE 17. PCA on leaf trait variables (nutrients phosphorus, mangnesium, calcium, potassium,
manganese (g.kg-1), LMA (g.m²), Cellulose, lignin and phenosl (%), based on means per plot. The two
principal components accounted for 51.2% of the variation data. The PCA included all eight nutrient
addition treatments of Control (red), Nitrogen (N) (yellow), Phosphorus (P) (light green), Cation (dark
green), N+P (light blue), N+Cation (turquoise blue), P+Cation (lilac), N+P+Cation (pink).
40
FIGURE 18. Simple correlation between concentration of the nutrients in insect frass biomass
input (g.ha-1) and leaf litter biomass input (g.ha-1) of potassium (a), magnesium (b), calcium (c) and
phosphorus (d) and figure e is the correlation between calcium (g.kg-1) and phenols.
41
TABLE 3. Table of variables in treatments control. It contains the type of material, unit, number
of repetitions (N), mean value, standard deviation (SD), standard error (SE), and confidence Interval (CI).
Variable Mean SD SE CI 95% N Nutrient concentration in litter (g kg-1) P 0.19 0.018 0.009 0.029 4
Ca 4.56 1.27 0.637 2.028 4
Mg 1.77 0.25 0.125 0.399 4
K 1.06 0.136 0.068 0.216 4
Mn 0.09 0.01 0.005 0.016 4
Nutrient concentration in frass (g kg-1) P 0.63 0.13 0.066 0.211 4
Ca 5.13 0.93 0.47 1.485 4
Mg 1.83 0.07 0.034 0.109 4
K 2.415 0.68 0.34 1.083 4
Mn 0.06 0.004 0.002
0.006 4
Nutrient input from litter (g ha-1) P 27.74 1.82 0.912 2.904 4
Ca 680.09 239.4 119.7 380.94 4
Mg 262.09 50.904 25.45 80.99 4
K 155.72 21.68 10.84 34.504 4
Mn 13.46 2.84 1.42 4.522 4
Nutrient input from frass (g ha-1) P 0.16 0.06 0.03 0.092 4
Ca 1.36 0.605 0.303 0.963 4
Mg 0.47 0.126 0.063 0.200 4
K 0.59 0.032 0.016 0.051 4
Mn 0.015 0.004 0.002 0.007 4
Concentration of defense
compounds in litter (%) Cellulose 27.94 2.17 1.085 3.453 4
Lignin 42.7 3.10 1.551 4.935 4
Phenols 0.01 0.002 0.001 0.003 4
LH (%) 8.11 2.63 1.315 4.183 4
LMM_Bh (g kg-1) 1.26 0.75
4
LMA (g.m²) 181.52 70.20 4
Litter biomass input (kg ha-1) 147.27 15.82 7.909 25.169 4
Frass biomass input (kg ha-1) 0.26 0.078 0.039 0.124 4
TABLE 4. Table of variables in all treatments. It contains the type of leaf nutrient input, unit
(kg.ha-1yr-1), number of repetitions (N), mean value, standard deviation (SD), standard error (SE).
0.374 1.192
35.102 111.710
42
0.353 4
1.76 4
N P (kg.ha-1yr-1)
CONTROL
CATIONS
N
N+CATIONS
N+P
N+P+CATIONS
P
1.75
1.83
1.87
1.56
2.49
2.23
2.24
0.0717
0.477
0.372
0.420
0.085
0.501
0.474
0.036
0.238
0.186
0.210
0.043
0.250
0.237
4
4
4
4
4
4
4
P+CATIONS 2.12 0.706
Mg (kg.ha-1yr-1) CONTROL 16.6 3.52 CATIONS 18.9 2.70 1.35 4
N 17.2 2.40 1.20 4
N+CATIONS 18.1 3.34 1.67 4
N+P 27.5 3.00 1.50 4
N+P+CATIONS 18.2 3.72 1.86 4
P 23.4 5.24 2.62 4
P+CATIONS 17.3 5.29 2.64 4
Ca (kg.ha-1yr-1) CONTROL 43.0 15.1 7.54 4
CATIONS 44.4 8.1 4.05 4
N 43.6 12.7 6.34 4
N+CATIONS 50.0 5.20 2.60 4 N+P 68.3 15.7 7.85 4
N+P+CATIONS 45.3 12.5 6.23 4
P 56.1 21.8 10.9 4
K (kg.ha-1yr-1)
P+CATIONS
CONTROL
37.0
9.83
7.88
1.15
3.94 4
0.57 4
CATIONS 11.8 4.94 2.47 4
N 10.4 2.67 1.33 4
N+CATIONS 11.4 1.45 0.72 4
N+P 13.5 4.06 2.03 4
N+P+CATIONS 10.0 1.81 0.91 4
P 10.0 3.29 1.65 4
P+CATIONS 10.7 2.46 1.23 4
Treatment Mean SD SE
43
6. Discussion
6.1. Short-term responses to fertilization
We investigated the short-term responses of litterfall production, nutrient inputs,
and herbivory to nutrient addition in a Central Amazonian forest. We expected that
nutrient additions would increase leaf quality resulting in an increase in leaf turnover
rates, or increased leaf litterfall biomass. The leaf litter nutrient inputs (except K) had a
strong significant response to the addition of P treatments, in the first three months
postfertilization. In contrast, leaf litterfall biomass had a trend for increased biomass with
the combined N + P treatment over the initial eight months post fertilization. With the
increase of leaf litterfall nutrient concentrations, we expected an overall increase in leaf
quality traits, such as a decrease in leaf mass per unit area (LMA) and defense compounds
(greater palatability). In contrast to our expectations, we found that LMA increased
significantly with added nitrogen and cations but defense compounds (lignin, cellulose
and phenols) had no response to nutrient additions. Together, this suggests that leaf litter
nutrient inputs are highly flexible in this low-nutrient tropical forest, whereas biomass and
defense traits are more stable with longer lag times before changes are found.
We expected herbivory to increase with foliar nutrient content and overall greater
palatability, but herbivory, as percentage leaf area lost, did not differ among the
treatments. We did find increased leaf mass missing (g.kg-1) with added nitrogen and
cations, but this response disappeared when scaling to hectare level. To verify herbivore
responses to nutrient addition we also quantified insect frass biomass and nutrient content,
expecting that we could find an influence of nutrient addition. Similar to our ecosystem
level herbivory estimates, cumulative insect frass did not differ between the treatments
and control. However, there was a strong response of insect frass nutrient input with the
addition of phosphorus. Together, the initial leaf litterfall and herbivore suggests that they
had a rapid initial responses to nutrient addition.
6.2. Effect of nutrient addition in litterfall production
Fine litterfall productivity (mean value 9.12±6.33 Mg ha-1 year-1) at our low
phosphorus site was similar to other low nutrient tropical sites (Mirmanto et al 1999,
Chave et al. 2010). The fine litterfall in this experiment had mean values in treatment N+P
(10.73±8.81 Mg ha-1 year-1) slightly above the rate found in different sites across
Amazonian lowland and montane sites of 8.61±1.91 Mg ha-1 year-1 (Chave et al., 2010;
Vitousek, 1984; Heineman et al., 2015). Similarly, leaf litterfall production in our
experimental site (5.73±0.46 (N) to 7.59±0.82 (N+P) Mg ha-1 year-1) was slightly higher
than the overall average of 5.6 Mg ha-1 year-1 for terra firme forests across the Amazon
(Chave et al. 2010).
44
Seasonality of litterfall production followed the pattern expected for the Brazilian
Amazon region of Manaus (Wu et al., 2016). Peak litterfall production was the first week
of september in all treatments, which follows patterns of other tropical forests with peak
production during the dry season (Heineman et al.,2015). Plants produce a bigger flush of
young leaves during the dry season, coinciding with lower damage rates by herbivores
(Albert et al.2018). This might be considered a defense mechanism as well, because even
though there are more palatable leaves, the higher amount of leaves might satiate the
insects enough to inflict significantly less damage to trees than when leaves are produced
in wet season (Aide, 1988;1991, Coley 1983). This suggests that leaf production occurring
in the beginning of dry season may have a slight temporal escape from herbivores, as it
also coincides with the period of lowest insect abundance, lowering damage rates in the
dry season that (Wolda, 1978; Aide, 1988; 1991).
The cumulative leaf litterfall (FIGURE 9
